Comment on “Scaling of Atmosphere and Ocean Temperature Correlations in Observations and Climate Models”

A Comment on the Letter by Klaus Fraedrich and Richard Blender, Phys. Rev. Lett. 90, 108501 (2003). The authors of the Letter offer a Reply.

Figure 1

Results of DFA2 for the daily maximum temperature records of 18 inner continental stations (○) and two coastline stations (×): (a) 1 Urumchi, 2 Tomsk, 3 Atbasar, 4 Chita, 5 Olekminsk, 6 Horog, 7 Swerdlowsk, 8 Surgut, 9 Jakutsk, 10 Aleksandrovsk; (b) 1 Huron, 2 Academy, 3 Cheyenne, 4 Gothenburg, 5 Gunnison, 6 Spokane, 7 Winnemucca, 8 Pendleton, 9 Tuscon, 10 New York. The estimated minimum distance to the oceans is written in parentheses. The scale of $F(s)$ is arbitrary. The straight lines in the curves represent the best fits between $s=150$ and 2500. For each curve, the variance of the slope is about $0.01$. The line at the bottom has a slope of $0.15$, corresponding to $\alpha =0.65$. The maximum $s$ value in each curve is below one quarter of the length of the corresponding record.

Figure 2

The scaling exponent $\alpha$ (obtained from the slopes in Fig. 1) as a function of the distance to the oceans.